At present, a laser pointer is an interactive device commonly used in a projection screen display system. A laser emitted by the laser pointer projects a laser point onto the screen, and control over a projection device can be effectively implemented by detecting, using a camera, a movement of the laser point on the screen.
As a projected area increases, it is difficult for a single camera to cover an entire screen. Multiple cameras are used to capture an image on the screen, and each of the cameras is responsible for capturing an image in one area on the screen. First, image fusion is performed on an overlapping area captured by different cameras; then laser point detection is performed on an entire frame to determine a location of a laser point; finally, information about the location of the laser point is output by means of coordinate transformation.
However, each time an image is captured, image fusion needs to be performed first, and then laser point detection is performed on an entire frame. This results in extremely poor real-time performance in an entire process of acquiring a location of a laser point.